Fusion approaches: tokamak, stellarator and laser

What has each fusion approach demonstrated, and what gap stands between it and power on a grid?

Tokamak, stellarator and laser fusion report their headline results in different units, so the figures are shown side by side with their conditions rather than ranked. None of the sources reports a fusion plant sending net electricity to a grid, and each row names a different main gap.

Fusion approaches: tokamak, stellarator and laser. Each cell shows a value, how well it is evidenced, and its sources.
Option	Headline demonstrated resultFacility, year and measured fusion energy in megajoules (MJ) for a pulse or shot, or pulse length in seconds (s) for magnetic devices with a duration record.	How confinement worksThe physical method used to hold the fuel while it fuses. Descriptive, no units.	Main engineering gapThe main thing not yet shown in a working device, with the figure that shows it, such as shot rate in shots per second or shots per day.	Next major milestoneThe next announced step, with its date where a source gives one. Dates are announced targets, not guarantees.
Tokamak	JET, UK: final tritium campaign (DTE3), announced 8 February 2024: 69 MJ of fusion energy in one pulse of about 5 seconds. Research machine, not electricity.Demonstrated[2]	Strong magnetic fields hold a doughnut-shaped plasma in a ring. A large electric current in the plasma helps create the twisted field.Demonstrated[2][4]	Plant engineering is left to later work. The ARC overview (June 2026) says engineering and economic analysis are outside its scope, and treats blanket and vacuum vessel details as baseline assumptions.Proposed[6]	ITER research operation in 2034 and deuterium-tritium operation in 2039, per the July 2024 schedule. The deuterium-tritium start is four years later than the earlier plan.Proposed[7]
Stellarator	Wendelstein 7-X, Germany, 22 May 2025: a new peak in long-pulse triple product, sustained for 43 seconds. The source gives no numeric value and does not describe the result as peer reviewed.Reported, unverified[3]	70 superconducting coils make the magnetic field that holds the plasma. The twist comes from the coil design rather than mainly from a plasma current.Demonstrated[4]	Pulse length: 43 seconds reported. IPP says the pellet method is directly relevant to future reactors and can potentially extend pulses to several minutes. That is not yet shown.Reported, unverified[3]	IPP says its pellet method can potentially extend plasma pulses to several minutes. The source names no campaign or date.Proposed[3]
Laser (inertial) confinement	National Ignition Facility, US, 7 April 2025: 8.6 MJ of fusion yield from 2.08 MJ of laser energy on target (target gain 4.13). Single shot.Demonstrated[1]	Lasers strike a hohlraum, and the X-rays it produces implode a small fuel capsule. The fuel's own inertia holds it together long enough for fusion to happen within a single shot.Demonstrated[1][5]	Shot rate: the National Ignition Facility fires one or two shots a day, according to the SLAC interview. The same interview gives about 10 shots per second as the rate needed for commercial fusion.Reported, unverified[5]	Further ignition shots at higher laser energy, as LLNL expects. The source gives no shot-rate target or date.Proposed[1]
Power on the grid	National Ignition Facility, 5 December 2022: 3.15 MJ of fusion energy from 2.05 MJ on target. This is energy in a laser shot, not grid electricity.Demonstrated[1]	No reliable public figure	The 400 MW net electric output in the ARC overview is a design target, not a demonstrated output. The paper says engineering and economic analysis are outside its scope.Proposed[6]	US DOE aims to support fusion pilot plants and commercial fusion power in the mid-2030s. This is a government goal, contingent on future appropriations.Proposed[8]

Fusion approaches: tokamak, stellarator and laser. Each cell shows a value, how well it is evidenced, and its sources.

Option

Headline demonstrated resultFacility, year and measured fusion energy in megajoules (MJ) for a pulse or shot, or pulse length in seconds (s) for magnetic devices with a duration record.

How confinement worksThe physical method used to hold the fuel while it fuses. Descriptive, no units.

Main engineering gapThe main thing not yet shown in a working device, with the figure that shows it, such as shot rate in shots per second or shots per day.

Next major milestoneThe next announced step, with its date where a source gives one. Dates are announced targets, not guarantees.

Tokamak

JET, UK: final tritium campaign (DTE3), announced 8 February 2024: 69 MJ of fusion energy in one pulse of about 5 seconds. Research machine, not electricity.Demonstrated[2]

Strong magnetic fields hold a doughnut-shaped plasma in a ring. A large electric current in the plasma helps create the twisted field.Demonstrated[2][4]

Plant engineering is left to later work. The ARC overview (June 2026) says engineering and economic analysis are outside its scope, and treats blanket and vacuum vessel details as baseline assumptions.Proposed[6]

ITER research operation in 2034 and deuterium-tritium operation in 2039, per the July 2024 schedule. The deuterium-tritium start is four years later than the earlier plan.Proposed[7]

Stellarator

Wendelstein 7-X, Germany, 22 May 2025: a new peak in long-pulse triple product, sustained for 43 seconds. The source gives no numeric value and does not describe the result as peer reviewed.Reported, unverified[3]

70 superconducting coils make the magnetic field that holds the plasma. The twist comes from the coil design rather than mainly from a plasma current.Demonstrated[4]

Pulse length: 43 seconds reported. IPP says the pellet method is directly relevant to future reactors and can potentially extend pulses to several minutes. That is not yet shown.Reported, unverified[3]

IPP says its pellet method can potentially extend plasma pulses to several minutes. The source names no campaign or date.Proposed[3]

Laser (inertial) confinement

National Ignition Facility, US, 7 April 2025: 8.6 MJ of fusion yield from 2.08 MJ of laser energy on target (target gain 4.13). Single shot.Demonstrated[1]

Lasers strike a hohlraum, and the X-rays it produces implode a small fuel capsule. The fuel's own inertia holds it together long enough for fusion to happen within a single shot.Demonstrated[1][5]

Shot rate: the National Ignition Facility fires one or two shots a day, according to the SLAC interview. The same interview gives about 10 shots per second as the rate needed for commercial fusion.Reported, unverified[5]

Further ignition shots at higher laser energy, as LLNL expects. The source gives no shot-rate target or date.Proposed[1]

Power on the grid

National Ignition Facility, 5 December 2022: 3.15 MJ of fusion energy from 2.05 MJ on target. This is energy in a laser shot, not grid electricity.Demonstrated[1]

No reliable public figure

The 400 MW net electric output in the ARC overview is a design target, not a demonstrated output. The paper says engineering and economic analysis are outside its scope.Proposed[6]

US DOE aims to support fusion pilot plants and commercial fusion power in the mid-2030s. This is a government goal, contingent on future appropriations.Proposed[8]

What the table shows

The National Ignition Facility shot of 7 April 2025 delivered 8.6 MJ from 2.08 MJ on target, and the facility fires one or two shots a day, according to the SLAC interview.
The JET figure of 69 MJ is for one pulse of about 5 seconds at a research facility. The Wendelstein 7-X figure is for 43 seconds and is not described as peer reviewed.
The tokamak row's main gap is plant engineering, the stellarator row's is pulse length, and the laser row's is shot rate.
No row shows a fusion plant sending net electricity to a grid. The grid row records the pilot-plant goal as a mid-2030s government aim.

Read with care

Measures differ. Laser figures are fusion energy against laser energy on the target, while JET and Wendelstein 7-X figures come from single pulses at research machines.
Figures are from announcements or single campaigns. The Wendelstein 7-X source has a 30 June 2025 update, and its JET comparison uses unpublished data.
Target gain compares fusion energy with laser energy on target. It does not count the electricity used to run the laser or the wider facility, so it is not net electricity.
Dates and targets are announced plans. ITER's schedule has already moved, and DOE says its roadmap does not commit it to funding levels.

Sources

other
Achieving Fusion Ignition
Lawrence Livermore National Laboratory, 2026. Accessed 21 Oct 2026.
government report
Fusion research facility JET's final tritium experiments yield new energy record
UK Atomic Energy Authority, 8 Feb 2024. Accessed 21 Oct 2026.
other
Wendelstein 7-X sets new performance records in fusion research [UPDATE]
Max Planck Institute for Plasma Physics, 3 Jun 2025. Accessed 21 Oct 2026.
other
Preparation for operation has begun
Martin Trinkaus. Helmholtz Association, 28 May 2014. Accessed 21 Oct 2026.
news
What will it take to bring fusion energy to the US power grid?
Erin Woodward. SLAC National Accelerator Laboratory, 7 May 2025. Accessed 21 Oct 2026.
peer reviewed paper
Overview of the physics basis for the ARC fusion power plant
Jon C. Hillesheim, Alexander J. Creely, Thomas H. Eich et al.. Journal of Plasma Physics, 4 Jun 2026. Accessed 21 Oct 2026.
other
The implications of the new ITER schedule
Max Planck Institute for Plasma Physics, 3 Jul 2024. Accessed 21 Oct 2026.
government report
Energy Department Releases Finalized Fusion Science and Technology Roadmap to Accelerate Commercial Fusion Power
US Department of Energy, 9 Jun 2026. Accessed 21 Oct 2026.